Femtosecond laser system
A detailed schematic of the femtosecond laser facility
is shown in Figure 1. A femtosecond mode-locked seed beam of 14.5 nm bandwidth,
pulse energies in the nanojoule range and repetition rate of 80 MHz is
emitted from a Ti:sapphire oscillator pumped by a diode laser. A pulsed
Nd:YLF operating at repetition rate of 1 kHz pumps the seed beam through
a regenerative amplifier. Using the chirped pulse amplification technique,
ultra-short pulses are generated with a FWHM pulse width of about 83 fs,
800 nm wavelength and 1 mJ maximum pulse energy.
Time resolved microscopy on semiconductors and metals using pump and probe technique
The pump and probe technique is a widely used experimental
method for observing phenomena happening within a 100 picosecond time period
(1 picosecond=10-12 second). For instance, chemists want to
know how the atoms in a molecule react to one another within 100 femtoseconds
(1 femtosecond=10-15 second). Physicists strive to elucidate
energy transfer mechanisms with the femtosecond laser pulse. The technique
itself involves splitting a short pulse laser beam into two beams, a pump
beam and a probe beam. The pump beam induces material changes after it
hits in the target material, and the probe beam monitors this transformation
at different time delays. The short pulse laser, which has up to several
tens of femtoseconds time resolution, is used because of higher time resolution.
The pump and probe technique can exploit the energy transfer
mechanism during irradiation of a ultrashort laser beam onto the target
material. Furthermore, thorough knowledge of the short-pulse laser interaction
with the target material is essential for controlling the resulting modification
of the micro-sized target structure.
During the present experiment, single pulses were used
at fixed pulse energy of 0.35 mJ on the target surface. This energy corresponds
to a peak power of about 4.3 GW. Furthermore, the following list shows
important optical components involved in the pump and probe beam paths,
as shown in Figure 2.
1. Non-Linear Crystal (NLC): doubles the frequency of an input beam.
2. Half and quarter waveplates (l/2, l/4): changes beam polarization.
3. Polarizing BeamSplitter (PBS): transmits or reflects a particularly
polarized beam.
4. Dichroic mirror (DM): selectively reflects or transmits a certain
beam depending on its wavelength.
The pump beam (wavelength=800 nm) path (purple line) is
relatively simple. The 90% reflected beam at the beam-splitter (BS) is
relayed to the delay stage and transferred to the DM. The delay stage is
installed to set the time difference between the pump and probe beam. This
pump beam is transmitted through the DM to heat up the sample.
The 10 % portion of the fundamental 800 nm beam serves
to generate a frequency-doubled (wavelength=400 nm) probe beam (blue line)
when going through the NLC. This probe beam is horizontally (P) polarized
before hitting the NLC. The NLC changes the polarization to the vertical
polarization (S). The half-waveplate (l/2) then changes the polarization
back to P so that it can pass through the polarizing beam splitter (PBS),
which transmits P-polarized light and reflects S-polarized light. This
S-polarized beam is converted to circularly polarized light by the quarter
waveplate (l/4). The dichroic mirror (DM) reflects the frequency-doubled
but transmits the fundamental beam.
The probe beam arrives at the sample specified delay time
after the irradiation by the pump beam. Then, the reflected probe beam
is directed through the quarter waveplate so that its polarization is converted
to S from the circular polarization. Thus, it can be reflected at the PBS.
The probe beam in this particular setup is interpreted as an image by the
CCD camera. The computer analyzes the image to obtain useful information
such as surface deformation, plasma formation, material ejection, etc.
The following figures show a sequence of surface images for varying
time delays between 0.2 ps and 90 ps under atmospheric gas pressure and
1 mtorr, respectively.
Time-of-flight mass spectroscopy
For the TOF measurement, the sample was mounted on a rotational
feedthrough in a vacuum chamber of 10-7 torr base pressure.
A fused silica spherical plano-convex lens (f = 250 mm) was employed to
focus the laser pulses onto the sample surface at a 50° angle of incidence
with respect to the normal direction. The repetition rate of the laser
was set at 3 Hz. A pulse generator (Stanford Research, DG535) was used
to trigger the laser and a 250 MHz digitizing oscilloscope (HP, Model 54510A).
The laser ablated ions drift through an 80 cm long field-free vacuum tube.
A microsphere plate detector (El-Mul Technologies) measured the ion TOF
spectra, which were used to determine the velocity distribution of the
ions. An instrument control program written in Labview (National Instruments)
handled data acquisition and processing. Each spectrum was obtained by
accumulating 300 sets of single shot data. The energies of the ions were
verified by setting repelling voltages at an electrode plate placed in
front of the detector. A Wiley-MacLaren type mass spectrometer (Wiley and
McLaren, 1955) was built for mass analysis by applying pulsing extraction
field to the positive ions before they reach the drift tube.
This figure shows typical time-of-flight spectra of laser-ablated titanium
ions.
The laser-induced plasma was studied by emission spectroscopy. The laser pulses were focused onto the sample via the same plano-convex lens at approximately normal incidence. An ICCD camera (Princeton Instruments, ITE/576BR) was used to capture the plasma emission images. The ICCD camera was equipped with a Nikon 105mm UV lens, of good transmission over the entire spectral range of the system from 250 to 800 nm. The emission spectrum of the plume was acquired for identifying the ablated species. A 150 mm UV lens projected the plume on a quartz optical fiber, which directed the emission into a monochromator (McPherson, Model 2035). The ICCD camera was mounted at the exit slit of the monochromator, which then effectively became a spectrometer.
This figure shows emission spectrum of laser-induced titanium plume.
Femtosecond pulsed laser recrystallization of a-Si film on quartz substrate
It has been thought that femtosecond pulsed laser recrystallization is not so desirable because the pulse duration is so short for crystallization. In fact, we believe that there is no normal melting phase. The recrystallization mechanism is going to be solid phase crystallization. A recent result of recrystallization is shown below.
Figure: Femtosecond pulsed laser recrystallization (the right figure represent the boundary of two different regime)
Click on the following movie to see the recrystallization dynamics and and the change of relative reflectivity.
Micromachining workstation for MEMS device (microcooler)
Ultra-short laser pulses impart extremely high intensities
and provide precise laser ablation (material ejection) thresholds at substantially
reduced laser energy densities. These advantages introduce precise control
of working materials and subsequently low thermal damges. A microcooler
device has a lot of holes and grooves that must be machined precisely.
The current micromachining workstation for the microcooler device realizes
a good control of finding spots and grooves and that of laser firing by
using LabVIEW interface.
The morphology, depth and volume of craters produced by
various number of laser shots both in moderate vacuum and at ambient pressure,
were measured by a white light interferometric microscope (Zygo, NewView
200) and a scanning electron microscope (SEM). A high-speed shutter (Vincent
Associates, Uniblitz LS6Z2) was used to control the number of shots that
hit on the fresh surface of the samples.
This figure shows the surface profile of a laser-ablated silicon crater
as measured by Zygo microscope.
Plasma study was done considering the pre-pulse effect on femtosecond laser micromachining. The following figure shows a typical plasma generated by surface emitted electrons.
